Tio₂/SiO₂/Fe₃O₄ magnetic nanoparticles synthesis and application in methyl orange UV photocatalytic removal

One of the most significant current discussions in the field of wastewater treatment is to find the most efficient method to convert the pollution from industrial wastewater into less harmful species (Luyten et al. 2013; Krzywicka & Kwarciak-Kozłowska 2014). Industrial wastewaters are sources for variety of contaminant species which should be eliminated from the stream through one of the physical (Kesari et al. 2011), chemical and biological (Beltran-Heredia et al. 2000) methods as well as coupled processes (Mozia et al. 2016). Releasing these wastewaters without further treatment will no doubt affect our environment and especially limited available water resources (Samaei et al. 2018). In one way or another, any contamination or pollution in one resource will eventually, if not immediately, affect the other. Contamination of groundwater resources by pathogenic organisms, inorganic and organic contaminants is going to be one of the most important issues for human being (Pham et al. 2019). Considering the lifetime and the release potential of the non-aqueous phase liquid (NAPL) contamination, preventing water sources from being contaminated by this type of pollution can be identified as the most challenging one.

Chemical (Fu & Wang 2011) and physical (Chu et al. 2019) methods including flocculation (Sadri Moghaddam et al. 2010), adsorption (Zhang et al. 2012; Pham et al. 2017, 2018), absorption (Chouchene et al. 2010), precipitation (Irawan et al. 2011), membrane separation (Le-Clech et al. 2006) and affinity membrane separation (Managheb et al. 2021) suffers from solid wastes formation, which necessitate an additional disposal or regeneration process. On the other hand, non-biodegradable pollutions limit the biological degradation processes (Basha et al. 2009).

Advanced oxidation processes (Aplin & Waite 2000; Andreozzi et al. 2004) including, sulfate radicals-based AOPs (SR-AOPs) (Xiao et al. 2020; Li et al. 2021), UV irradiation (Xu et al. 2018), photocatalysis (Matos et al. 2014; Khoshnavazi et al. 2016), ozonation (Tichonovas et al. 2017), Fenton oxidation (Sekaran et al. 2014; Han et al. 2017), electrochemical oxidation (Trinh et al. 2012; Woisetschläger et al. 2013; Nihei et al. 2018), catalytic oxidation (Do et al. 2018), and ultrasound assisted processes (Amaniampong et al. 2017, 2018, 2019) are much more efficient compared to above mentioned methods in elimination of contaminations from wastewater (Oturan & Aaron 2014). Through all these techniques, radicals are formed which are highly active in oxidation of organic pollutions in contaminated water (Deng & Zhao 2015). Among all, photocatalytic technique is privileged due to high efficiency, low cost, and stability of the catalysts against different corrosion sources.

Titanium oxide is a photosensitive semiconductor can be announced as one of the most attractive catalyst which can be used in degradation of organic compounds, toxic pollutions in waste and air (Liu et al. 2006; Han et al. 2009; Chen et al. 2012; Hua et al. 2012; Da Dalt et al. 2013; Bozkurt Çırak et al. 2019). TiO₂ UV irradiation exposure results in formation of valance band holes and conduction band electrons which provides redox media within which toxic substances and organic pollutions can be easily degraded into simple molecules such as carbon dioxide and water which are not toxic and do not even need any further disposal or regeneration steps (Han et al. 2009; Da Dalt et al. 2013).

As mentioned earlier, TiO₂ is one of the most active substances in photocatalytic wastewater treatment (Zhang et al. 2017). However, it should be considered that anatase phase is much more active compared with rutile phase (Luttrell et al. 2014; Matos et al. 2014). To increase TiO₂ photocatalytic activity further, small and nano size particles are favorable, which enhances photocatalytic activity by increasing specific surface area (Xu et al. 1999), but it causes separation issues in catalyst recovery from treated water. Immobilization of TiO₂ using supports results in drastic reduction in specific surface area and, also, it is not that successful due to loose connection and interaction between TiO₂ and the support (Shan et al. 2010). To benefit high specific surface area of nano-size TiO₂ photocatalyst particles besides efficient catalyst recovery, supported titania on nanoparticle size magnetic powders can be considered as an alternative choice (Qiao et al. 2003; Životský et al. 2015; Chen & Fu 2018).

Using magnetic core of iron oxide (Fe₃O₄) as support for TiO₂ adversely affects activity while enhancing practical aspects of catalyst recovery and separation issues (Yao et al. 2015; Guo et al. 2017). To overcome the adverse effects of using Fe₃O₄, presence of a thin layer of SiO₂ between Fe₃O₄ magnetic core and photocatalytically active TiO₂ shell is inevitable (Gad-Allah et al. 2007; Yuan et al. 2012; Chi et al. 2013; Larumbe et al. 2014; Wang et al. 2017b). This middle neutral layer of SiO₂ reduces the interaction between the core and the shell, while selective supporting of anatase phase rather than rutile amplify the enhancing impact (Yu et al. 2011; Santiago et al. 2018).

In this study, TiO₂/SiO₂/Fe₃O₄ nanoparticles were prepared and used in photocatalytic removal of methyl orange. Methyl orange was selected for this study for two main reasons. The first reason is the resistance of methyl orange to UV radiation. This means that in the presence of UV radiation, this dye, unlike dyes such as red 51, is not photolysis. In other words, the only process responsible for the decomposition of methyl orange is the photocatalytic process, which takes place on the surface of the synthesized nanoparticles. The second reason is the widespread use of Methyl orange in the dyeing industries. Effects of catalyst loading, pH of the reaction media and aeration rate were studied. The response surface methodology was used to design experiments. The goals we are pursuing during this research are:

• 1.

  Investigation and modeling of direct and interactive effects of operating parameters of methyl orange photocatalytic degradation process.

• 2.

  Maximizing the performance of TiO₂/SiO₂/Fe₃O₄ magnetic catalysts in the process of photocatalytic removal of methyl orange from synthetic wastewater.

• 3.

  Complete removal of methyl orange organic contamination from a synthetic wastewater and complete separation of photocatalytic nanoparticles from processed wastewater

Fe₃O₄ with a purity of 98% and a diameter of 20 nm, ethanol (C₂H₅OH) with a purity of 99.9%, ammonium hydroxide (NH₄OH) with a purity of 0.25%, tetraethyl orthosilicate (SiC₈H₂₀O₄) 30% (TEOS) were used as received from Merck KGaA and titanium isopropoxide (Ti(OCH(CH₃)₂)₄) 97% was provided by SIGMA-ALDRICH and was used without any further processes. One molar solution of sulfuric acid and sodium hydroxide was also used to adjust the pH value of the reaction media.

Preparation of magnetic photocatalytic nanoparticles were performed in two stages. The first stage is preparation of electromagnetic Fe₃O₄ cores covered in SiO₂ shells. The second stage in the preparation of photocatalytic particles is to coat the SiO₂/Fe₃O₄ nanoparticles with the TiO₂ photo-active layers.

If uncoated Fe₃O₄ is used as a support for TiO₂, the performance of the prepared photocatalyst will be reduced due to the interactions between them. Therefore, a thin layer of SiO₂ should be coated on the Fe₃O₄ magnetic core to reduce the interaction between the magnetic core and the photocatalytically active TiO₂ shell (Wang et al. 2017b). Precipitation method was used to coat SiO₂ particles of Fe₃O₄. 15 mL of ammonium hydroxide were added to 0.3 g of Fe₃O₄ in 87.2 mL of ethanol (Dayana et al. 2019). The solution was diluted to 200 mL using deionized water. The synthesis solution was placed in an ultrasonic bath for 10 min. Afterward, 10 mL of an aqueous solution containing 7 mL of tetraethyl orthosilicate was added dropwise to the synthesis solution and the final solution was aged under very slow shaking for 18 h. Upon completion, the particles were separated from the solution using magnetic force, washed with ethanol and dried in an electric oven heated to 70 °C (Absalan & Nikazar 2016).

SiO₂/Fe₃O₄ nanoparticles were coated by titanium dioxide using sol-gel method. Fe₃O₄ particles covered by SiO₂ were dispersed in 10 mL of ethanol and placed in an ultrasonic bath for 10 min. After adding 15 mL ethanol and 12 mL of deionized water, a solution of 1 mL of titanium isopropoxide in 5 mL of ethanol were added. The synthesis solution was slowly shacked for 16 h and the final solution was washed with deionized water, dried at 70 °C for 10 h and calcined at 450 °C for 3 h (Yao et al. 2015).

X-ray diffraction (XRD) patterns of SiO₂/Fe₃O₄ and TiO₂/SiO₂/Fe₃O₄ nanoparticles were obtained using Equinox 3000, Inel France, using Co-Kα X-ray tube with an input voltage of 40 kV while 2θ varied from 20° to 80° with scan rate of 1°/min, and 2θ intervals of 0.02°.

AIS2100, seron technology, was used to obtain SEM micrographs and study morphology of the prepared SiO₂/Fe₃O₄ core support and TiO₂/SiO₂/Fe₃O₄ photocatalyst.

Attenuated reflection Fourier transform infrared (ATR FT-IR) spectra of SiO₂/Fe₃O₄ nanoparticles were obtained using VERTEX 70, Bruker, USA, in the wavenumber range of 400–4,000 cm⁻¹.

Textural properties of the synthesized nanoparticle were studied using surface area analyzer (TriStar-II-Series, Micromeritics Instrument Corporation, USA). The specific surface area of the synthesized photocatalyst nanoparticles was evaluated using Brunauer–Emmett–Teller (BET) method. Barrett–Joyner–Hacienda (BJH) method was used to calculate the average pore size and the pore size distribution using adsorption and desorption branch of the isotherm. Before the adsorption–desorption studies, the nanoparticles were degassed at 300 °C for 3 h.

To determine point of zero charge (PZC) of the catalyst, the salt addition technique was used (Mullet et al. 1999; Bakatula et al. 2018). The measurement was started by adding 0.25 g of the prepared photocatalyst to solutions with the same ionic strength but variety of pH values. The prepared photocatalyst was added to 50.0 mL of 0.1 M NaNO₃ solution. The pH values of the solutions were adjusted to 3, 4, 5, 5.5, 6, 6.5, 7, 8 and 9 using solutions of 0.1 M HNO₃ and 0.1 M NaOH. The pH value of each solution was measured before and after 24 h moderate mixing for a day using a shaker incubator and values are denoted as pH_i and pH_f, respectively. The plot of pH variation (ΔpH) during this process against pH_i was used to obtain the PZC.

TiO₂/SiO₂/Fe₃O₄ nanoparticles were used in photocatalytic degradation of methyl orange. For each experiment, 200 mL of the feed containing 100 ppm of methyl orange were charged to the reactor. The required amount of TiO₂/SiO₂/Fe₃O₄ nanoparticle was determined by experimental design, were added to the reaction media, the pH of which was adjusted to the value dictated by the experimental design. Immediately after adding the catalyst, the reactor was covered from light, mixing and aeration with specified flow rate determined by experimental design was started in presence of the UV lamp (15 watt). Figure 1 schematically illustrates the apparatus.

If a photocatalyst is used in the treatment of municipal or industrial wastewaters, especially when using nanoparticles as the photocatalyst, the need for particle recycling will be both economically and environmentally crucial. In this research, separation of the magnetic photocatalyst nanoparticles was performed using a 1.3 Tesla magnet.

The separated nanoparticles were reused for photocatalytic process. After several (6) steps of the photocatalytic decomposition process, the photocatalyst regeneration process were performed. For this purpose, the nanoparticles were immersed in distilled water and irradiated with UV light for 12 h.

After regeneration, photocatalytic activity evaluation was performed again.

Response surface methodology (RSM) was used to develop a model and analyze the obtained results. Effects of three potentially important parameters in photocatalytic performance of TiO₂/SiO₂/Fe₃O₄ nanoparticles were studied and the conditions which resulted in maximum catalytic activity was determined by optimization of the response surface. RSM was used in combination with a fractional design method of Box-Behnken which required fewer experimental sets compared with common fractional design methods while optimization accuracy is the same (Asadi et al. 2019).

β₀, β_i, β_ii, and β_ij are constants which were determined in the second-order model. ε is a random error.

The differences in significance of independent parameters was identified using analysis of variance (ANOVA), sum of squares and F statistics (Amini et al. 2008; Bashir et al. 2010; Asadi et al. 2016, 2018; Iravaninia & Rowshanzamir 2016).

The XRD patterns for the SiO₂/Fe₃O₄ cores, and the TiO₂/SiO₂/Fe₃O₄ composite, are shown in Figure 2. As evidenced, SiO₂ has not been shown to a specific peak, which is due to the presence of SiO₂ as an amorphous phase.

As shown in Figure 2, the XRD peaks confirms the presence of the anatase TiO₂ phase as well as Fe₃O₄, indicating the survival of the Fe₃O₄ nanoparticle crystal structure after TiO₂ addition. All corresponding peaks to (101), (004), (200), (105) and (211) planes are present in the XRD pattern of the synthesized photocatalyst (the red pattern), according to which, prepared TiO₂ powder can be indexed as anatase phase with standard JCPDS card number of 21-1272 (Li et al. 2014). Obtained peaks are in good agreement with the ones reported in other research articles (Dao et al. 2018). Clearly, all corresponding peaks to (220), (311), (222), (400), (422), (511), (440) and (620) planes which are attributed to the cubic Fe₃O₄ (JCPDS Card No. 19-629) are present in the XRD pattern of the Fe₃O₄/SiO₂ support (blue pattern) (Ma et al. 2015; Wang et al. 2017a; Setoodeh et al. 2019).

To confirm the presence of all the mentioned elements including Fe, Si, Ti and O in the synthesized photocatalyst nanoparticles, elemental analysis was performed on the sample using energy dispersive X-ray spectroscopy (EDS) analysis. Figure 3 presents the obtained EDAX spectrum, which confirms the presence of all these elements. Results revealed that the atomic ratio of Fe : Si : Ti is equal to 0.42 : 2.08 : 1.00.

Nanoparticles size and surface morphology were investigated by SEM, results are depicted in Figure 4. The pure and non-coated Fe₃O₄ nanoparticles micrograph is shown in Figure 4(a)–4(c) illustrate the synthesized SiO₂/Fe₃O₄ and TiO₂/SiO₂/Fe₃O₄ nanoparticles morphology, respectively. Considering Figure 4(c), the spherical and almost identical structure of the nanoparticles can be confirmed.

The specific surface area of the prepared photocatalyst nanoparticles was measured to be 56 m²/g. The pore volume and the average pore size of the sample were 0.12 cm³/g and 9.4 nm, respectively. The nitrogen gas adsorption and desorption isotherm are shown in Figure 5. Clearly, the isotherm can be described by the standard IUPAC classification of adsorption–desorption isotherm of Type III. The synthesized photocatalyst particles give rise to H₃ hysteresis which states the presence of slit-shaped pores (Al-Othman 2012). Besides, samples with this type of adsorption–desorption isotherm are typically formed from non-rigid aggregates (Al-Othman 2012).

Figure 6 shows the FT-IR spectroscopy for SiO₂/Fe₃O₄ nanoparticles.

Presence of the absorption peaks at 575 cm⁻¹ and 1,110 cm⁻¹ in the FT-IR spectrum; which could be respectively assigned to a group of bands typical of stretching vibrations of Fe–O bonds (in the range of 567–578 cm⁻¹) (Wang et al. 1998) and asymmetric stretching vibrations of Si–O(Si) (around 1,100 cm⁻¹) (Adamczyk & Rokita 2016). Additionally, bands correspond to the vibrations of OH⁻ groups are observed at around 3,430 cm⁻¹ as well as those of molecular water at 1,630 cm⁻¹ (Adamczyk & Rokita 2016). The presence of bands at and around 2,922 cm⁻¹ and 1,450 cm⁻¹ were respectively attributed to stretching in C–H (Ta et al. 2016) and C–H₂ (Xu et al. 2013) groups of organic compounds, such as TEOS, which were used in the synthesis procedure. It can be inferred that silica coating on the surface of the Fe₃O₄ nanoparticles is successful. FTIR results confirm successful coating of SiO₂ on Fe₃O₄ nanoparticles surface.

The designed experiments and experimentally obtained data as well as the predicted response values, are shown in Table 1.

Optimal conditions for maximization of the photocatalytic activity of the synthesized TiO₂ catalysts were determined through RSM by analyzing the response surface obtained from mathematical modeling of the experimental data. Statistical surveys of linear, two-factor-interaction, quadratic and cubic models were performed and the results are summarized in Table 2. Based on the Box-Cox technique, there is no need for any kind of transformation.

Fit quality is represented by R², which is the ratio of justified variations by the model to the total variation and a minimum value of 0.8 is an index for a good model. On this basis, the developed model is accurate. The contours shown in Figure 7 represents the methyl orange removal as a function of catalytic loading and pH, considering the constant value of 2.5 L/min for aeration.

As clearly seen in Figure 7, the methyl orange photocatalytic removal increases by increasing catalyst loading up to 2,000 ppm. This could be attributed to the active site abundance resulted by higher loadings, which consequently enhances absorption of higher percentage of photons transmitted to the reaction media as well as reactant molecules.

By decreasing the pH of the reaction media, the methyl orange photocatalytic removal rate increases and reaches the highest value in the pH of 3. The pH of the reaction media has complex effects on the photocatalytic oxidation reaction rate. The optimum pH value depends on the type of pollutant and the point of zero charge (PZC) of the catalyst, which was measured to be 5.9, as shown in Figure 8. This is in good agreement with the results reported in the literature (Khedr et al. 2019). At pH = pH_PZC, the surface is neutral, while positively and negatively charged photocatalyst surfaces can be considered at pH < pH_PZC and pH > pH_PZC, respectively (Singh et al. 2015). The adsorption of specific compound on the catalyst surface and its decomposition rate depends on both of their charge status. Clearly, pH of the solution governs the adsorption rate which affects the following photocatalytic reaction rate.

Regarding methyl orange, it should be noted that with the hydrolysis of the sodium sulfate group in its chemical structure, a methyl orange anion is formed. At different pHs, two different structures are conceivable for methyl orange, both of which have negative charges. It is necessary to mention the pK_a for methyl orange, which is equal to 3.7, in which case the neutral state will be maintained. At pHs above 5.9 (pH > pH_PZC) the TiO₂ surface is negatively charged. Due to the negative charge of the methyl orange anion, it is normal for electrostatic repulsion to occur between the surface of the photocatalyst and the dye anion. This phenomenon affects the rate of dye anion adsorption on the surface of the photocatalyst and thus reduces the rate of decomposition. At pH values below pH_PZC, the surface charge of the photocatalyst particles becomes positive. As the pH decreases from the pH_PZC value to about 3.7, the dissociation of the dye molecules decreases and therefore, as mentioned, the neutral state is established. However, with a further decrease in pH to about 3.1, and the formation of methyl orange anion due to dye dissociation, and formation of a negatively charged anion, the adsorption of the dye anion on the photocatalyst surface intensifies, and consequently suitable conditions for the photocatalytic reaction are provided. In addition, the formation of HO₂^• radicals at low pHs, in addition to compensating for the low concentration of OH^• radicals, intensifies the reactions (Bouanimba et al. 2015).

High aeration rates, increases the amount of oxygen and consequently increases the efficiency of OH° radical production and prevents the recombination reaction. Besides, it improves the reaction media agitation which reduces diffusion resistance for reactants. These two simultaneously enhance the photocatalytic reaction and reduce the remained unreacted methyl orange in the reaction zone.

The analysis of variance (ANOVA) results for linear model is summarized in Table 3. The model is significant based on the calculated F-Value of 39.85 and P-Value of 2.19 × 10⁻⁷. The larger F-Value and small P-Value calculated for catalytic loading (parameter A) compared with those of other parameters, stating that this parameter is the most significant one.

To validate the proposed model, the accuracy of predicted values should be checked. The comparison between the experimental and predicted values for methyl orange removal was used to study the accuracy of the developed model. Figure 9 shows the linear relationship between experimental and predicted values for methyl orange removal which confirms the accuracy of the developed model. Distribution of the experimental data and the corresponding predicted values around the line within a narrow range clearly illustrates the model validity.

Experimental results revealed that it is possible to completely remove methyl orange. To achieve the optimal conditions under which this happens, model optimization was performed. In order to achieve the maximum methyl orange removal, the independent variables of catalyst loading, pH and aeration should be respectively adjusted to 2,000 ppm, 3 and 2.5 L/min, which results in 101.64% methyl orange removal, predicted by the model. However, to achieve 100% methyl orange removal, the developed model suggests a range of independent parameters, which are demonstrated in Figure 10.

To validate the proposed model, a catalyst was prepared with 2,000 ppm active phase loading. This catalyst was used in photocatalytic removal of methyl orange at pH of 3 while the reaction mixture was aerated with flow rate of 2.5 L/min. The results revealed that complete removal of methyl orange was achieved under this condition. Considering the obtained results, it could be concluded that the proposed model is valid.

A 1.3 Tesla magnet was used to separate the magnetic photocatalyst nanoparticles. Figure 11 shows the effects of magnetic field on TiO₂/SiO₂/Fe₃O₄ magnetic nanoparticles only after 1 min. It is clear that the separation of nanoparticles is well done.

Figure 12 shows changes in the photocatalytic activity of the nanoparticles due to separation and reuse. As can be seen in Figure 12, after six applications of magnetic photocatalytic particles, a slight decrease in photocatalytic activity was observed. However, after the regeneration operation, in which the catalyst is immersed in distilled water for 12 h under UV light, the catalyst returns to its original state. After several regeneration operations, a decrease in the efficiency of photocatalyst activity is observed. This reduction can be attributed to the blockage of the active photocatalytic pores due to the deposition of insensitive materials to the radiation process or the loss of nanoparticles in each recycling (Beydoun et al. 2000).

In this paper, we investigate the photocatalytic performance of TiO₂/SiO₂/Fe₃O₄ nanoparticles using a response surface methodology. ANOVA results revealed that photocatalytic performance is governed by the direct effects of catalyst loading, pH of reaction media and aeration flow rate. XRD, EDAX and FTIR results confirmed the successful composite particles formation, which exhibited specific surface area, pore volume and average pore size of 56 m²/g, 0.12 cm³/g and 9.4 nm, respectively. The adsorption–desorption isotherm follows Type III-H₃ hysteresis stating the presence of slit-shaped pores. The PZC of the catalyst was measured to be 5.9. The results of catalyst performance evaluation revealed that the optimum condition for maximization of the photocatalytic methyl orange removal is catalyst loading of 2,000 ppm, reaction media pH value of 3, and aeration rate of 2.5 L/min. TiO₂/SiO₂/Fe₃O₄ nanoparticles were well recovered under the magnetic field.

The TiO₂ photocatalytic activity was enhanced by optimizing process conditions through developing a statistical model for the performance of the photocatalytic process. By establishing optimum process conditions, it is possible to completely remove methyl orange. In addition, the magnetic properties of the photocatalyst particles make them completely separable from the processed water stream. Findings suggest that all studied factors are effective and the proposed TiO₂/SiO₂/Fe₃O₄ nanoparticles are an alternative photocatalyst to treat wastewaters and reduce their environmental impacts.

We would like to offer our special thanks and sincere gratitude to deceased Professor Morteza Sohrabi (Department of Chemical Engineering, Amirkabir University of Technology), who, although no longer with us, continues to inspire by his manner of thought. He guided us in every aspect of this research.

All relevant data are included in the paper or its Supplementary Information.